VERMEULE_3FMT
05/17/2006 09:21:16 AM
Essay
Political Constraints on Supreme Court
Reform
Adrian Vermeule†
Many proposals to reform the rules of the Supreme Court
game are currently under discussion.1 Which of these proposals
lie within the politically feasible set, and which are ruled out by
political constraints? In what follows, I will sketch the shape of
those constraints and describe the main political mechanisms
that produce them. I use the failure of Roosevelt’s courtpacking plan in 1937 as a running example, supplemented by
comparisons with the flurry of reform plans—mostly unsuccessful—offered during Reconstruction. The main thesis is that reform of the Court requires political conditions that have a selfnegating tendency. The very conditions that produce demand
for structural reform of the Court also tend to produce counterforces that block the movement for reform. The point is not of
course that structural reform is impossible, in the sense that it
is always ruled out by political constraints. In particular cases,
the demand for reform may be just strong enough, and the
counterforces produced by that demand just weak enough, that
a reform proposal can slip through. Yet reform cannot be predicted in advance or relied upon; it is systematically unlikely to
occur. The stronger the movement for reform, the higher the
obstacles that must be surmounted.
† Bernard D. Meltzer Professor of Law, University of Chicago. Thanks
to Dennis Hutchinson, Gerry Rosenberg, Adam Samaha, Fred Schauer, Cass
Sunstein, and Mark Tushnet for helpful comments and conversations. Thanks
to Sean Heikkila for helpful research assistance. The Russell J. Parsons Faculty Research Fund provided generous support.
1. See, e.g., Paul D. Carrington & Roger C. Cramton, The Supreme Court
Renewal Act: A Return to Basic Principles, in REFORMING THE COURT: TERM
LIMITS FOR SUPREME COURT JUSTICES 467 (Roger C. Cramton & Paul D. Carrington eds., 2006).
1154
VERMEULE_3FMT
2006]
05/17/2006 09:21:16 AM
POLITICAL CONSTRAINTS ON REFORM
1155
My principal interest is in detailing the mechanisms that
produce political constraints on Supreme Court reform, regardless of how tight those constraints turn out to be. However, I
will also suggest, without attempting to provide systematic evidence, that the constraints are in fact quite restrictive. The ash
heap of history is piled high with reform proposals that have
attracted no supporters (other than those who formulated
them), attracted academic supporters but no political backers,
or attracted political backers but no popular following. Almost
all ideas for Supreme Court reform die in committee, literally
or metaphorically. The constitutional and statutory rules governing the Court—the number of its members, their terms of
tenure, the voting and quorum rules that govern their actions,
and so on—have in most cases remained unchanged, at least
since Reconstruction, and in some cases since the first Judiciary Act of 1789.2 Not everything has held constant—the switch
to discretionary certiorari jurisdiction in 1925 is a salient example3—but in the broad, structural reform of the Court is exceedingly rare. All else equal, the higher the stakes of a reform
proposal, the more opposition it will generate and the less
likely it is to succeed. The reform proposals that do succeed,
conversely, are likely to be of marginal importance, at least
when compared to the ambitious model of the court-packing
plan.
Part I defines “reform” as structural change in the constitutional and statutory rules that govern the Supreme Court
game, as opposed to substitution of new players for old ones
(through appointments) or new behavior by old players under
the old rules (a “switch in time”). Parts II through V turn to the
mechanisms that constrain reform by provoking counterforces
to the reform movement. Although some of these mechanisms
apply to institutional reform generally, some apply only to reform of the Court. Court reform both partakes in the general
difficulty of institutional reform and presents additional difficulties of its own.
Part II discusses the problem of multidimensional politics.
The large, national coalitions necessary for Supreme Court reform will typically be assembled on other issue dimensions and
will fracture when judicial reform comes to the fore. Part III
2. Act of Sept. 24, 1789, ch. 20, 1 Stat. 73.
3. See Kenneth W. Starr, The Supreme Court and Its Shrinking Docket:
The Ghost of William Howard Taft, 90 MINN. L. REV. 1363, 1364 (2006).
VERMEULE_3FMT
1156
05/17/2006 09:21:16 AM
MINNESOTA LAW REVIEW
[90:1154
considers the problem of the optimal majority. Reform movements must steer between Scylla and Charybdis: a majority
that is too small will be blocked at the vetogates of the legislative process, while a majority that is too large will provoke a
backlash spurred by fears of tyranny. Part IV discusses the basic trade-off between impartiality and motivation. Structural
reforms adopted behind a veil of uncertainty will be and seem
impartial, but in general, no politically influential group will be
motivated to support them. Conversely, proposals that produce
short-term benefits for particular groups will attract motivated
supporters but will also provoke opposition. Part V suggests
that political crisis is both a precondition for and an obstacle to
reform of the Court. In a brief conclusion, I consider the relevance of political constraints from the standpoints of both analysts and advocates of reform. Although analysts should consider political constraints, advocates of reform should not.
I. “REFORM”
We may define reform both by reference to paradigm cases
and at the conceptual level. I will take as the paradigm of reform Roosevelt’s 1937 court-packing bill. So far as relevant
here, the proposal would have added one Justice, up to a total
membership of fifteen, for each Justice over the age of 70 who
had served ten years and who did not retire within six months
of his 70th birthday.4 Roosevelt had carried forty-eight states in
the 1936 election and commanded filibuster-proof majorities in
both the House and Senate (although we will see that the Democratic coalition would fracture along the fault line of Supreme Court reform).5 My main thesis about the episode will be
that the very conditions that produced such obvious potential
for reforming the Court also produced the political constraints
that blocked reform.
At the conceptual level, I will generalize from the courtpacking example to stipulate that “reform” means a proposal
for change in the rules of the Supreme Court game. This definition makes reform synonymous with structural reform, including the number of Justices, their tenure, voting rules, and so
forth. It excludes both (1) a substitution of new players for old
4. See WILLIAM E. LEUCHTENBURG, THE SUPREME COURT REBORN: THE
CONSTITUTIONAL REVOLUTION IN THE AGE OF ROOSEVELT 134 (1995).
5. Id. at 158–59.
VERMEULE_3FMT
2006]
05/17/2006 09:21:16 AM
POLITICAL CONSTRAINTS ON REFORM
1157
players through the appointments process and (2) a change in
the actions that the old players take under the old rules.
This definition has the consequence that the indirect effects of the court-packing plan did not produce reform in the
structural sense I have indicated. Roosevelt eventually obtained no less than seven Supreme Court appointments, producing a cadre of like-minded Justices.6 For reasons discussed
below, however, this does not count as reform. Furthermore, if
the selection of new players does not count, a change in the behavior of the old players does not count either. Suppose that
the threat of court packing—or the anticipation of some threat
of that kind—produced a “switch in time,” in which Justice
Roberts changed his vote to uphold politically controversial
economic and social legislation.7 (Here, I am bracketing a set of
historical controversies over whether there was any such
switch and whether, if there was, it was caused by the courtpacking plan).8 This change in the actions taken by old Justices
under the old rules is not structural reform; it is tacking with
the prevailing political winds. Besides the court-packing plan,
another example involves the proposal, floated during both Reconstruction and the New Deal, to require a two-thirds vote of
the Justices to invalidate legislation on constitutional grounds.9
During Reconstruction, the proposal lacked a critical mass of
support, in part because the Court ducked many of the central
constitutional issues posed by Reconstruction legislation and
thus vented away the growing pressure for reform.10
6. See id. at 220.
7. For a discussion on the “switch in time,” see generally id. at 142–47.
8. On the question whether a switch in time occurred, see BARRY CUSHMAN, RETHINKING THE NEW DEAL COURT: THE STRUCTURE OF A CONSTITUTIONAL REVOLUTION 45–105 (1998); G. EDWARD WHITE, THE CONSTITUTION
AND THE NEW DEAL 11–163 (2000). Another view is that there was a switch,
but it predated the court-packing plan and was caused by the Justices’ anticipation of a threat of constitutional amendment following the 1936 election. See
Rafael Gely & Pablo T. Spiller, The Political Economy of Supreme Court Constitutional Decisions: The Case of Roosevelt’s Court-Packing Plan, 12 INT’L
REV. L. & ECON. 45, 53–65 (1992). Presumably, one might also hypothesize
that the switch predated the court-packing plan but was caused by the Justices’ anticipation of that plan or another plan with similar effects.
9. See DAVID P. CURRIE, THE CONSTITUTION IN THE SUPREME COURT:
THE FIRST HUNDRED YEARS 307 & n.144 (1985); David E. Kyvig, The Road
Not Taken: FDR, the Supreme Court, and Constitutional Amendment, 104
POL. SCI. Q. 463, 470–71 (1989).
10. See CURRIE, supra note 9, at 307 & n.144.
VERMEULE_3FMT
1158
05/17/2006 09:21:16 AM
MINNESOTA LAW REVIEW
[90:1154
Why define reform so narrowly? If Roosevelt obtained
much of what he wanted through the switch in time and
through appointments, why classify the court-packing plan as a
failure? Roosevelt himself said that although he lost the courtpacking battle, he won the broader war.11 From an even
broader perspective, however, the problem is that changes not
amounting to structural reform are only a temporary palliative.
As long as the rules themselves remain unchanged, later periods can see a recurrence of the problems that motivated reform
in the first place. Even if the outcome of the court-packing fight
was good for Roosevelt or the New Deal in the short run,12 it
may have been bad for the polity in the long run.
This point holds whether one considers the actual aim of
the court-packing plan or its nominal aim. Below, I will discuss
the divergence between the plan’s actual aim and its nominal
aim, which produced the contemporary perception that the plan
was a thinly disguised ploy for packing the Court with ideologically compatible Justices. As for the actual aim of the courtpacking plan—clearing away judicial obstruction to New Deal
policies—some of Roosevelt’s advisers “argued that Justices
who could switch so easily in his favor could just as easily jump
back once the pressure was off.”13 Although no such switch
back took place in the period, the advisers were right in a larger time frame. The absence of a formal New Deal amendment
to the Constitution permits or at least encourages the Court to
flirt with retro-restrictive interpretations of the Commerce
Clause, in line with a broader originalist movement that some
say aims to reinstate the pre-1937 “Constitution in Exile.”14 As
for the nominal goal of the court-packing plan—reducing the
fraction of Justices whose sell-by date has passed—the direct
effect of the Roosevelt plan would have been to place a structural cap on the fraction of the Court’s membership that was 70
11. See Kyvig, supra note 9, at 466.
12. For discussion and references on whether the outcome of the fight was
in Roosevelt’s interests (short-term or long-term), see generally Jamie L. Carson & Benjamin A. Kleinerman, A Switch in Time Saves Nine: Institutions,
Strategic Actors, and FDR’s Court-Packing Plan, 113 PUB. CHOICE 301 (2002).
13. LEUCHTENBURG, supra note 4, at 144.
14. This phrase stems from Douglas H. Ginsburg, Delegation Running
Riot, 18 REGULATION 83, 84 (1995). For debate about whether any such “Constitution in Exile” movement even exists, see Cass Sunstein & Randy Barnett,
Constitution In Exile?, LEGAL AFF., May 2, 2005, http://www.legalaffairs
.org/webexclusive/dc_printerfriendly.msp?id=41.
VERMEULE_3FMT
2006]
05/17/2006 09:21:16 AM
POLITICAL CONSTRAINTS ON REFORM
1159
years old or more, an issue that has arisen again in recent
years.15
The narrow definition of reform makes clear the first sense
in which Supreme Court reform tends to be a self-negating enterprise. The conditions that produce a real threat of reform
will also, for the same reasons, tend to produce a switch in time
that reduces the demand for reform. The threat of reform tightens the political constraints on the Justices, but a switch in
time can buy off the threat. Of course, if it is clear ex ante that
reform will be blocked by political constraints, then the threat
of reform is not credible from the Justices’ standpoint. From the
standpoint of reformers, however, the Justices’ ability to tack
as the winds set against them is itself another political constraint on reform. The judicial volte-face saps the political demand for structural change.
II. THE PROBLEM OF MULTIDIMENSIONAL POLITICS
I begin in this section with the well-known problem that
majority coalitions organized along other lines may fracture
once Supreme Court reform is put on the table. Given huge majorities in both Houses of Congress and a landslide victory in
1936,16 how could Roosevelt have failed to secure enactment of
the court-packing plan? In a standard account, the New Deal
coalition that had been created in and for the elections of 1932
and 1936 splintered badly over the court-packing issue.17 In the
latter election, Roosevelt had campaigned on a largely backward-looking platform that asked for a national referendum on
the first wave of New Deal reform.18 The issue of judicial reform was not raised.19 The court-packing plan of 1937 then split
the Roosevelt coalition by introducing a new issue dimension.20
Southern and western Democrats who supported the New Deal
on other grounds were opposed to tampering with the Court.21
15. See, e.g., Steven G. Calabresi & James Lindgren, Term Limits for the
Supreme Court: Life Tenure Reconsidered, in REFORMING THE COURT: TERM
LIMITS FOR SUPREME COURT JUSTICES, supra note 1, at 15; David J. Garrow,
Mental Decrepitude on the U.S. Supreme Court: The Historical Case for a 28th
Amendment, 67 U. CHI. L. REV. 995 (2000).
16. See LEUCHTENBURG, supra note 4, at 132.
17. Id. at 158–59.
18. Id. at 107.
19. Id.
20. Id. at 158–59.
21. Id.
VERMEULE_3FMT
1160
05/17/2006 09:21:16 AM
MINNESOTA LAW REVIEW
[90:1154
In general, conservative Democrats and Republicans found a
common cause in opposing court packing.22
Stipulating to the validity of this account, the important
point is that it captures more than the happenstance of 1936
and 1937. Generally speaking, reform of the Supreme Court is
rarely a central organizing issue in national elections. The
Court is just not that important when viewed in a larger perspective and compared to the bedrock issues of wages, taxes,
housing, and other policies that do determine national political
elections. Occasionally, as in 1968, the Court will come to be
loosely associated with a national political issue, such as
“crime,” but political platforms will be vague about what to do
about the Court in order to paper over differences within the
majority coalition. By and large, national coalitions must be organized on other issue dimensions, which means that they will
be inherently unstable when reform of the Supreme Court becomes the central issue. The conditions that make Supreme
Court reform possible—that a large national majority be organized—also tend to militate against the success of the reform
proposal. The majority coalition will tend to come unglued
when faced with the reform issue itself.
So far, I have mentioned two possibilities. First, straightforwardly, a coalition may be organized around an issue and
have intense preferences concerning that issue, as with the
Roosevelt coalition and the New Deal economic program. Second, as in the case of the court-packing plan, a coalition that is
organized around one issue may fracture when faced with a different issue as to which coalition members have intense preferences in opposite directions. Thirdly, however, there is another
important possibility: a coalition may be organized around one
issue and, on a separate issue of court reform, hold only weak
preferences or even be indifferent. In this scenario, reform is
possible because the coalition’s leaders, elected on another program, may seize the opportunity to implement reform if they
happen to favor it, without fracturing their political base.23
However, this is generally possible only with reform proposals that provoke no strong opposition, which will also tend
to be low-stakes proposals for marginal reforms. Where the
governing coalition is indifferent about the reform because the
benefit to them is small, but there is strong opposition because
22. Id.
23. Thanks to Mark Tushnet for emphasizing this possibility.
VERMEULE_3FMT
2006]
05/17/2006 09:21:16 AM
POLITICAL CONSTRAINTS ON REFORM
1161
the costs to others are high, leaders will be unable to mobilize
enough support to secure enactment. More generally, the
higher the stakes in a reform proposal, the more opposition it
will tend to provoke. To be sure, reform can and does occur, but
this tendency means that most of the reforms that do succeed
will be marginal accomplishments. Consider in this regard the
history of jurisdictional reform of the Court. Despite many proposals for stripping the Court of jurisdiction over one policy
area or another, no important jurisdiction-stripping measures
have been enacted since Reconstruction.24 The important jurisdictional changes that have occurred, such as the abolition of
most of the Court’s mandatory appellate jurisdiction in 1925,25
have uniformly expanded the Justices’ power and discretion.
III. THE PROBLEM OF THE OPTIMAL MAJORITY
In this section, I turn to a more speculative account of the
failure of the court-packing plan. It is obvious that a reform
movement may fail because it commands no majority or commands a majority that is too small; the foregoing account,
based on the fracturing of the Democratic coalition, suggests
that the rump of Roosevelt supporters was too small to overcome the opposition. The problem of an excessively small majority, however, is matched by an equal and opposite problem:
the majority favoring reform may also be too large.
The first risk, that of an insufficient majority, is familiar.
At the level of national public opinion, even a clear majority
may fail to translate into a majority in national lawmaking institutions, given certain distributions of voters in a first-pastthe-post electoral system. Even if there exists a majority coalition in Congress that is organized on the issue of judicial re-
24. For an overview of jurisdiction-stripping proposals since Reconstruction, see ERWIN CHEMERINSKY, FEDERAL JURISDICTION § 3.1 (4th ed. 2003)
(noting that “few jurisdictional restrictions have been adopted thus far in
American history”). It has been argued that, contrary to the conventional wisdom, jurisdiction stripping is common. See Dawn M. Chutkow, Jurisdiction
Stripping: Ideology, Institutional Concerns, and Congressional Control of the
Court 2 (Oct. 19, 2005) (unpublished working paper), available at http://papers
.ssrn.com/sol3/papers.cfm?abstract_id=828906. This thesis seems to depend
upon an overly broad definition of jurisdiction stripping. Although it is difficult
to be sure, the author’s definition seems to conflate jurisdiction stripping with
preclusion of judicial review in the first instance and seemingly does not distinguish between cases of genuine jurisdiction stripping and cases in which
Congress simply transferred jurisdiction from one court to another.
25. See Chutkow, supra note 24, at 2.
VERMEULE_3FMT
1162
05/17/2006 09:21:16 AM
MINNESOTA LAW REVIEW
[90:1154
form—thus bracketing the problem of multidimensional politics—vetogates in the legislative process, including bicameralism, congressional committees, and the filibuster, may allow
minorities on the reform issue to block change. The puzzle of
the court-packing episode, however, is that after the 1936 election, Roosevelt held the presidency and partisan majorities in
both Houses of Congress so large as to negate the threat of a
partisan filibuster.26 On paper, the vetogates had already been
unlocked, yet the reform plan failed ignominiously.27
Crucially, the very size of the Roosevelt majority itself
seems to have produced widespread public concern about the
court-packing plan. In this perception, it suddenly became clear
that Roosevelt effectively controlled the nonjudicial branches of
government, that the narrow Court majority was one of the few
remaining focal points for resistance to the New Deal program,
and that a risk of executive tyranny had arisen.28 In the general case, a majority that is too large may provoke a backlash
by creating fears of untrammeled power among those who distrust government in general or the executive in particular, and
by increasing the underlying distrust itself. As Senator Henry
Ashurst put it, “[e]ven many people who believe in President
Roosevelt . . . were haunted by the terrible fear that some future President might, by suddenly enlarging the Supreme
Court, suppress free speech, free assembly, and invade other
Constitutional guarantees of citizens.”29
On this view, proposing a plan to enlarge the number of
Justices played a crucial role in sharpening public concern
about executive despotism generally and Roosevelt’s intentions
in particular. Two factors may explain why this was so. The
first is the “normative power of the factual.”30 The number of
Justices had been set at nine since 1869, when the Reconstruction Congress increased the number from seven in order to give
Ulysses S. Grant extra appointments.31 Such longstanding
26. See LEUCHTENBURG, supra note 4, at 132.
27. Actually, there were two reform plans, both of which failed. See William E. Leuchtenburg, FDR’s Court-Packing Plan: A Second Life, A Second
Death, 1985 DUKE L.J. 673, 673. I touch upon the second plan in Part IV, infra.
28. See LEUCHTENBURG, supra note 4, at 137.
29. Michael Nelson, The President and the Court: Reinterpreting the
Court-Packing Episode of 1937, 103 POL. SCI. Q. 267, 276 (1988).
30. See LEUCHTENBURG, supra note 4, at 137–42.
31. Conversely, in 1866, Congress reduced the number of Justices from
ten to seven in order to deny appointments to President Andrew Johnson. See
VERMEULE_3FMT
2006]
05/17/2006 09:21:16 AM
POLITICAL CONSTRAINTS ON REFORM
1163
rules or invented traditions,32 whose dubious origin is usually
lost in the mists of time, often come to seem normatively significant. Although the number of Justices had been changed
several times before, many believed that the Constitution specified nine. One writer encountered an elderly lady who protested, “If nine judges were enough for George Washington,
they should be enough for President Roosevelt.”33
The precise mechanisms that generate the normative
power of the factual are as yet poorly understood; conjecturally,
there is some relationship to the endowment effect, under
which subjects value what they have more than what they
lack,34 or to the phenomenon of status-quo bias.35 The norms
that arise from institutional arrangements seen as longstanding, however, tend to support an inference that actors who contravene such norms are untrustworthy. Partially informed publics or voters may use the following heuristic: any political actor
who seeks to change the rules in the middle of the game is untrustworthy, presumptively motivated by partisan advantage or
a desire for unchecked power.
The second factor was a widespread perception that the
court-packing plan was a disingenuous proposal.36 Although
purportedly based on a concern about the competence of aging
judges, the plan was widely seen as a gambit to increase the
number of New Deal supporters on the Court.37 When it became clear that the initial reaction to the plan was negative,
Roosevelt tried to link the two issues, suggesting that antiquated judges would also have obsolete and socially harmful
views.38 By that time, however, the damage had been done.
ROBERT L. STERN ET AL., SUPREME COURT PRACTICE 2 n.7 (7th ed. 1993).
32. See Eric Hobsbawm, Introduction to THE INVENTION OF TRADITION 1,
1–4 (Eric Hobsbawm & Terence Ranger eds., 1983).
33. LEUCHTENBURG, supra note 4, at 139.
34. See, e.g., Daniel Kahneman & Amos Tversky, Prospect Theory: An
Analysis of Decision Under Risk, 47 ECONOMETRICA 263, 277–80 (1979).
35. See, e.g., William Samuelson & Richard Zeckhauser, Status Quo Bias
in Decision Making, 1 J. RISK & UNCERTAINTY 7, 8 (1988).
36. See LEUCHTENBURG, supra note 4, at 138.
37. See id.
38. The famous fireside radio chat of March 9, 1937, is an example. See
President Franklin D. Roosevelt, Fireside Radio Chat Defending the Plan to
“Pack” the Supreme Court (March 9, 1937), in FDR’S FIRESIDE CHATS 83, 90
(Russell D. Buhite & David W. Levy eds., 1992), available at http://www.hpol
.org/fdr/chat.
VERMEULE_3FMT
1164
05/17/2006 09:21:16 AM
MINNESOTA LAW REVIEW
[90:1154
This episode illustrates the dilemmas inherent in “the civilizing force of hypocrisy.”39 Those who were unsure about the
merits of the plan seem to have relied upon another political
heuristic: that disingenuity implies bad motives. On this view,
dictatorship was the real face behind the mask of the courtpacking plan. It does not follow, however, that the same voters
would have seen Roosevelt as well motivated had he openly
confessed to the goal of removing ideological opponents from
the Court; that goal could in turn be condemned as partisanship. Political actors are constrained to offer a public-regarding
justification for reform, one that does not map too obviously
onto their ideological views or partisan interests. Roosevelt
failed by offering a plan that was too transparently motivated
by other considerations, but sincerity about those other considerations would not have improved the situation.
Moreover, the constraint that one must offer a purportedly
neutral criterion for reform was far from toothless; it actually
caused Roosevelt to overshoot the mark set by his political interests. Needing at most four more loyalist votes to gain control
of the Court,40 Roosevelt offered a neutral principle, in the form
39. Jon Elster, Alchemies of the Mind: Transmutation and Misrepresentation, 3 LEGAL THEORY 133, 176 (1997).
40. In fact, there is a plausible case for picking any number from one to
four as the magic number needed to gain control; which answer is correct depends on how the question is specified. The simplistic argument for one is that
many of the salient decisions invalidating New Deal programs (and similar
state programs) before 1937 were issued by a vote of 5–4. See, e.g., Morehead
v. New York ex rel. Tipaldo, 298 U.S. 587, 618 (1936) (finding a New York
minimum wage law unconstitutional), overruled by Olsen v. Nebraska ex rel.
W. Reference & Bond Ass’n, 313 U.S. 236 (1941); R.R. Ret. Bd. v. Alton R.R.
Co., 295 U.S. 330, 374 (1935) (holding the Railroad Retirement Act unconstitutional). But the Court was divided into three camps, not two. The Four
Horsemen (Justices Van Devanter, McReynolds, Sutherland, and Butler) were
reliable anti-New Deal votes, whereas Justices Brandeis, Stone, and Cardozo
were reliable in the other direction. LEUCHTENBURG, supra note 4, at 132–33.
Justices Hughes and Roberts were, in the view of ardent New Dealers at least,
unreliable weather vanes. Id. at 133. Even if one of the Horsemen were replaced with a Roosevelt loyalist, a coalition of Hughes, Roberts, and the remaining three Horsemen could still defeat New Deal programs by a 5–4 vote.
Assuming a Court with a constant membership of nine, Roosevelt needed two
new appointments to be certain of success. However, if new appointments
were additions rather than replacements, no less than three appointments
would be necessary to ensure a 6–6 tie (which would become a 7–5 majority
whenever either of the two waverers joined the liberals), while four appointments would be necessary to ensure a 7–6 victory. Under any scenario, however, six new appointments would be excessive, so the criterion Roosevelt offered overshot the mark.
VERMEULE_3FMT
2006]
05/17/2006 09:21:16 AM
POLITICAL CONSTRAINTS ON REFORM
1165
of an age cutoff of 70,41 that would have given him no less than
six additional appointments. With an age cutoff of 75, only five
more seats would have been added,42 but that figure would
have made the plan appear even more obviously pretextual.
Because three Justices were clumped at age 75, it was not possible to propose a cutoff that would have given Roosevelt exactly four more appointments. Even were such a criterion possible, however, it would have been so precisely tailored to
Roosevelt’s interests as to be suspicious in its own right. The
overshooting plausibly contributed to the impression that Roosevelt sought dictatorial powers. On the other hand, the 70year figure was a neutral-seeming round number and had the
political advantage of using the civilizing force of hypocrisy as a
sword: Roosevelt’s most intractable opponent among the
Horsemen, Justice McReynolds, had at an earlier time himself
proposed a retirement age of 70 for all federal judges.43
The puzzle for this whole line of explanation is why the
same voters, or a decisive fraction of the same voters, who supported Roosevelt and congressional Democrats in the 1936 election would react against Roosevelt’s large majority after the
fact. The multidimensional character of coalition politics would
supply an answer to this, but we are assuming away that issue
here. Assuming, counterfactually, that the 1936 election had
been organized as a referendum on the courts, why would ex
ante supporters of the New Deal become opponents of Rooseveltian tyranny ex post?
I conjecture that the striking turnabout occurred because
electoral decisions are uncoordinated. Rational voters, conditional on having entered the voting booth, vote expressively
rather than instrumentally.44 (I bracket here the well-known
41. LEUCHTENBURG, supra note 4, at 134.
42. On February 5, 1937, the date of the plan’s formal introduction, see
CUSHMAN, supra note 8, at 11, the ages of the Justices were 80 (Brandeis); 77
(Van Devanter); 75 (Hughes, McReynolds, and Sutherland); 70 (Butler); 66
(Cardozo); 64 (Stone); and 61 (Roberts). See generally ALPHEUS THOMAS MASON, HARLAN FISKE STONE: PILLAR OF THE LAW 11 (1956); 3 THE JUSTICES OF
THE UNITED STATES SUPREME COURT: THEIR LIVES AND MAJOR OPINIONS
(Leon Friedman & Fred L. Israel eds., 1997). Ignoring Hughes and Roberts as
unpredictable, any cutoff less than 71 and greater than 66 would maximize
the difference between the number of conservatives covered (four) and the
number of liberals covered (one). However, a cutoff at age 70 also possessed
the other benefits discussed in text.
43. See William E. Leuchtenburg, The Origins of Franklin D. Roosevelt’s
“Court-Packing” Plan, 1966 SUP. CT. REV. 347, 391–92.
44. See GEOFFREY BRENNAN & LOREN LOMASKY, DEMOCRACY AND DECI-
VERMEULE_3FMT
1166
05/17/2006 09:21:16 AM
MINNESOTA LAW REVIEW
[90:1154
prior puzzle of why rational people would vote in the first
place.) Any particular individual voter or group of voters, if assured of being the decisive voter, might prefer a division of partisan power across national lawmaking institutions, including
the Supreme Court. Given that any individual voting decision
has effectively zero chance of being decisive, however, voters
will indulge expressive statements of loyalty—to Roosevelt or to
the Democratic party—that are effectively costless. Once appraised that other voters have done likewise, however, the very
voters who supported Roosevelt and the congressional Democrats seem to have become concerned that the collective outcome
of uncoordinated choices had produced total Rooseveltian
dominance of the national government.
The upshot of all this is that reform proposals need an optimal majority to succeed—one that is neither too large nor too
small. In the court-packing episode, the political conditions that
produced a sufficient congressional majority to unlock the vetogates and surmount the hurdles of the lawmaking process also
caused the reform movement to overshoot the mark, producing
a backlash animated by fear of executive tyranny. The very factors that made reform possible also produced counterforces that
blocked its accomplishment. Again, this is only to identify a
mechanism that tends to block reform, not an iron law that reform can never succeed. In some cases, reform proposals can
slip between Scylla and Charybdis; the Reconstruction-era
changes to the Court’s composition are an example.45 In general, however, it is much more likely that any given proposal
will be sucked into the whirlpool of vetogates or swallowed up
in a political backlash against an excessive majority.
IV. THE VEIL OF UNCERTAINTY, DELAY,
AND MOTIVATION
A standard idea in constitutional design is that less information can be better than more. Depriving political actors of
information about how to promote their narrow self-interest
will cause them to behave as though animated by impartial motives. The veil of ignorance,46 or uncertainty,47 subjects decision
SION: THE PURE THEORY OF ELECTORAL PREFERENCE
32–53 (1993).
45. In 1866, Congress reduced the number of Justices from ten to seven,
in order to deny appointments to President Andrew Johnson. See STERN ET
AL., supra note 31, at 2 n.7.
46. See JOHN RAWLS, A THEORY OF JUSTICE 118–23 (rev. ed. 1999).
VERMEULE_3FMT
2006]
05/17/2006 09:21:16 AM
POLITICAL CONSTRAINTS ON REFORM
1167
makers to uncertainty about the distribution of benefits and
burdens that will result from a decision. This distributive uncertainty can take one of two basic forms. In the Rawlsian veil
of ignorance, decision makers are placed under a constraint of
ignorance about their own identities and attributes.48 This introduces uncertainty by allowing the decision maker to know
the distributive consequences of a decision across citizens—call
them A and B—but denying the decision maker the knowledge
of whether she herself occupies A’s position or B’s position. The
more common and less radical version, however, is the veil of
uncertainty.49 Although the relevant decision makers know
their own current identities and interests, the veil introduces
uncertainty about who will reap the greater gains from the decision.50
How is the veil of uncertainty produced? One technique is
to enact reforms at a certain time that will not take effect until
a later time, perhaps years or even decades hence.51 Delay
takes advantage of a preexisting uncertainty, the inherent unpredictability of the decision makers’ long-term interests, that
would otherwise be overwhelmed by the incentives to focus on
short-term considerations.52 By confining the decision’s application to the remote future in which decision makers’ interests
are unpredictable, a delay rule ensures that the only period
current decision makers can affect is one that is, from their ex
ante standpoint, subject to a veil of uncertainty.53
By diluting current decision makers’ ability to assess
where their own interests lie, delayed enactments might be
thought to ease the path of reform. One might imagine a variant of Roosevelt’s proposal in which the legislation, although
47. See GEOFFREY BRENNAN & JAMES M. BUCHANAN, THE REASON OF
RULES: CONSTITUTIONAL POLITICAL ECONOMY 30 (1985); John C. Harsanyi,
Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of
Utility, 63 J. POL. ECON. 309 (1955).
48. See RAWLS, supra note 46, at 118–19.
49. Harsanyi, supra note 47.
50. See id.
51. For discussion of “delayed-effect” rules, see Ariel Porat & Omri Yadlin,
Promoting Consensus In Society Through Deferred-Implementation Agreements, 56 U. TORONTO L.J. (forthcoming 2006), available at http://law.bepress
.com/cgi/viewcontent.cgi?article=1022&context=taulwps; Adrian Vermeule,
Veil of Ignorance Rules in Constitutional Law, 111 YALE L.J. 399, 419–24
(2001).
52. Vermeule, supra note 51, at 419–20.
53. Id.
VERMEULE_3FMT
1168
05/17/2006 09:21:16 AM
MINNESOTA LAW REVIEW
[90:1154
enacted now, would not take effect until after the 1940 election.54 In the general case, delayed implementation might buy
off the opposition that would otherwise be offered by political
actors whose interests would be harmed by a proposal with
immediate effect. In a similar vein, Paul Carrington and Roger
Cramton’s recent proposal for reforming the Justices’ tenure to
a fixed term of years55 in effect delays the proposal’s onset date
by grandfathering currently sitting Justices. The grandfathering presumably buys off self-interested opposition on the part of
sitting Justices.56
In fact, however, successful delayed enactments are quite
rare. A major reason is the basic trade-off between impartiality
and motivation.57 Delay may buy off opposition, but it also
eliminates the short-term, self-interested benefit that would
accrue to the political actors who would otherwise support the
proposal and incur the costs of shepherding it through the legislative process. Delay replaces self-interested motivation with
impartial reason, but the latter motivation is frequently too
feeble to produce action; the pallid claims of reason rarely provoke the same degree of energy and activity as the prospect of
self-interested gain. In general, the high opportunity costs of
political action, constricted agenda space in Congress, sharp
limits on the amount of time the executive can spend on any
one project, the horizon of reelection, and the tendency to discount the future, all push political actors to rank projects by
the degree of benefit they produce in the near term. Projects
that will produce large collective benefits in the long run, but
whose distributive valence is uncertain, will be subordinated to
projects that produce larger factional benefits in the short run.
In the case of Roosevelt, a court-packing plan with a delayed
effective date would not have advanced Roosevelt’s immediate
54. In fact, Roosevelt’s second court-packing plan embodied a partial delay mechanism. It authorized the President to “appoint an additional Justice
per calendar year for each member of the Supreme Court who remained on the
bench after the age of seventy-five.” LEUCHTENBURG, supra note 4, at 148.
Given the distribution of the Justices’ ages, “the bill would empower [Roosevelt] to name four new Justices, as well as a Justice to fill the Van Devanter
vacancy, but the total of five could not be reached until the beginning of 1940.”
Id.
55. See Carrington & Cramton, supra note 1.
56. This is my interpretation of the proposal’s effect; I do not assert that
Carrington and Cramton intend this effect.
57. For other explanations, see Porat & Yadlin, supra note 51, manuscript
at 29–34.
VERMEULE_3FMT
2006]
05/17/2006 09:21:16 AM
POLITICAL CONSTRAINTS ON REFORM
1169
goal of clearing judicial obstruction to New Deal programs. In
the case of the Carrington and Cramton proposal, the safe prediction is that it will never be enacted. The delayed onset of the
proposal makes its political valence uncertain, which in turn
means that no current political actor will be strongly motivated
to shoulder the burdens of enactment. Political action oriented
toward the long run tends to occur only where the political valence of the proposal is clear ex ante, as in proposals for environmental regulation with long-term effects.
Overall, the trade-off between impartiality and motivation
creates another Scylla and Charybdis through which reform
proposals must slip. On the one hand, stipulating that a proposal should have a delayed effective date helps to remove the
taint of self-interested or partisan motives and avoids triggering the heuristic that one should not try to change the rules in
the middle of the game. On the other hand, delaying the proposal’s effect means that concrete problems of the here and
now, the sort of problems most likely to motivate reform, cannot be addressed. The delaying tactic makes reform possible by
creating an appearance of impartiality and buying off current
opposition, but the tactic also makes the reform less likely to be
proposed and pursued. Conversely, the short-term interest that
motivates reform itself tends to create self-limiting political
constraints. Proposals that produce short-term benefits for particular groups will attract motivated supporters but will also
provoke opposition.
V. CRISIS, REFORM, AND CONSTITUTIONAL
AMENDMENTS
In this section, I will briefly generalize some of the foregoing points. A standard idea is that political constraints limit institutional reform until some exogenous shock creates a crisis
that destabilizes extant institutions.58 The resulting uncertainty can promote reform, as previously discussed. Exogenous
shocks or crises, however, are a necessary but insufficient condition for reform. This is because crisis has two effects pulling
in opposite directions: crisis destabilizes institutions, but it also
tends to create new political constraints that shore up those institutions against change. As relevant here, times of crisis both
increase the demand for Supreme Court reform and also tend
58. See, e.g., THRAINN EGGERTSSON, IMPERFECT INSTITUTIONS:
TIES AND LIMITS OF REFORM 152–73 (2005).
POSSIBILI-
VERMEULE_3FMT
1170
05/17/2006 09:21:16 AM
MINNESOTA LAW REVIEW
[90:1154
to create political obstacles to that reform. For concreteness, I
focus on a dilemma that plagued Roosevelt: should reform of
the Court be attempted through the process of constitutional
amendment or through the legislative process? The latter path
better suits the conditions of urgency that give rise to the demand for reform yet also creates an appearance of partisan
manipulation. The former path seems more suitable for structural reform yet may take too long to be an effective response to
a crisis.
Roosevelt and his advisers ultimately chose the legislative
path, in large part because they feared that the amendment
process would unduly delay reform. The sense of national economic crisis that gave rise to the New Deal agenda also implied
that immediate action was necessary. As Attorney General
Homer Cummings said, “[t]he Administration cannot very well
let its social program bog down because of adverse Supreme
Court decisions, and, on the other hand, the delays incident to
amendment are rather appalling.”59 However, the decision to
proceed legislatively also created or strengthened opposition to
the substance of the reform proposal.
In 1937, opponents of Roosevelt’s proposal for statutory
court reform fell into three camps. Some flatly opposed the content of the proposal, whatever its legal form.60 Others said that
the proposal should not or could not be enacted legislatively,
but could and should be enacted as a constitutional amendment.61 Within the latter group, we may distinguish those who
held their position in bad faith from those who held it in good
faith. Roosevelt suspected pervasive bad faith, saying that “the
same forces which are now calling for the amendment process
would turn around and fight ratification on the simple ground
that they do not like the particular amendment adopted by the
Congress.”62 However, the bill was also opposed by some, such
as Senator George Norris, who genuinely favored the substance
59. Kyvig, supra note 9, at 476 (quoting Diary of Homer Cummings, Attorney Gen. (Nov. 15, 1936)). The delay might not have been as protracted as
the New Dealers feared, see id. at 479, but the fear was certainly reasonable in
light of the salient example of the 1924 Child Labor Amendment, which was
still unratified as of 1937 (and which eventually failed). See id. at 479 & n.68.
60. Cf. id. at 466 (discussing the battle that ensued after Roosevelt announced his plan to reorganize the judiciary).
61. See id. at 467.
62. Id. at 477 (quoting Letter from Franklin D. Roosevelt, President,
United States of America, to Charles Burlingham (Feb. 23, 1937)).
VERMEULE_3FMT
2006]
05/17/2006 09:21:16 AM
POLITICAL CONSTRAINTS ON REFORM
1171
of the proposal but who also genuinely thought that constitutional amendment was the proper path.63
The rationale for the last position was never clearly stated.
Although an amendment would be necessary if the bill were
unconstitutional, the arguments to that effect were quite weak
given the Reconstruction precedents in which Congress had
manipulated the number of Justices at will.64 In any event, the
principal concerns motivating good-faith opposition to the
statutory path seem to have been nonlegal. Norris worried that
anything short of an amendment would “plague our descendants” because “it does not strike permanently at the evil we
want to remedy.”65 As suggested above, this concern applies
even more strongly to nonstructural reform accomplished by
constraining the Justices to execute a switch in time. Another,
vaguer intuition seems to have been that it was inherently
more suitable to pursue structural reform of the judiciary by
amendment rather than by statute. The intuition, similar to
the heuristics described above, is that changing the rules of the
judicial game by legislation is an attack on the referee by one of
the players, and thus presumptively arises from partisan or
self-interested motivations.
Overall, the choice of legal instruments for judicial reform
presented Roosevelt and his advisers with a dilemma. On the
one hand, the New Dealers knew from personal experience that
“[d]emands for . . . constitutional revision tend to arise in times
of crisis in which waiting is an unaffordable luxury.”66 The crisis that produced the demand for reform also constrained the
choice between available paths to reform by ruling out the more
protracted amendment process. On the other hand, the statutory avenue itself provoked political opposition that would not
otherwise have arisen. That opposition came both from badfaith opponents who were enabled to argue for the amendment
path—and were thus given an extra arrow for their rhetorical
quivers—and also from good-faith opponents, whose preference
for the amendment path put them in unwilling coalition with
the bad-faith opponents. The very crisis that produced the de63. See id. at 480.
64. See, e.g., STERN ET AL., supra note 31, at 2 n.7 (discussing Congress’s
reduction of the number of Justices from ten to seven to deny appointments to
President Johnson).
65. Kyvig, supra note 9, at 480.
66. JON ELSTER, ULYSSES UNBOUND: STUDIES IN RATIONALITY, PRECOMMITMENT, AND CONSTRAINTS 145 (2000).
VERMEULE_3FMT
1172
05/17/2006 09:21:16 AM
MINNESOTA LAW REVIEW
[90:1154
mand for reform by constraining the feasible paths to reform
also produced endogenous opposition to the alternative path. In
this sense, the impetus to Supreme Court reform that arose out
of the New Deal crisis had a self-negating tendency.
Again, this conjunction of forces and counterforces is not
inevitable. Occasionally, reform will slip through because there
is an evident structural problem, because all concerned are uncertain about the future effects of the reform, and because no
group will lose much from current enactment. The constitutional amendments that have restructured presidential succession may count as an example.67 In general, however, such a
constellation of happy circumstances is both fortuitous and
rare. Structural reform, especially of the Court, is systematically unlikely and unreliable.
CONCLUSION
I have suggested a range of mechanisms that systematically tend to make Supreme Court reform a self-negating enterprise. The very conditions that give rise to the demand for
reform also tend to create political constraints on reform.
Whether valid or not, this thesis would be relevant from the
standpoint of the external analyst, such as the academic social
scientist. I hasten to add, however, that the shape of the politically feasible set is not obviously a relevant consideration for
practical advocates of reform (who may also happen to hold
academic posts). A plausible division of labor is that the reformer should deliberately ignore political feasibility; she
should simply propose first-best plans and programs and then
let politics itself filter the feasible from the infeasible. The reformer is typically an expert in the substantive area at hand,
not in politics, so self-censoring in light of the reformer’s estimate of political feasibility carries the risk that the reformer
will mistakenly filter out an ideal solution that would actually
have been enacted. By contrast, the opposite error—the proposal of ideal solutions that are politically infeasible—is solved
automatically by the operation of the political filter. For these
reasons, reform proposals such as that offered by Carrington
and Cramton are socially very valuable, even if, from the analyst’s standpoint, they are systematically likely to fail.
67. See U.S. CONST. amends. XX, XXV.